Type Ia Supernovae As Not-Quite-So-Standard Cosmological Candles

Shag writes "Type Ia supernovae are used as cosmological 'standard candles' to measure distance because of their strong similarity to one another. This has made possible, for example, the research into universal expansion that led to the Nobel-winning discovery of 'dark energy.' For years, astrophysicists believed white dwarves exploded when they accreted enough mass from companion stars to reach a limit of 1.38 times the mass of our Sun. A decade ago, the 'Champagne supernova' (SN 2003fg) was so bright astrophysicists concluded the limit had been exceeded by two white dwarves colliding. Now a new paper (PDF) from the Nearby Supernova Factory collaboration suggests that type Ia supernovae occur at a wider range of stellar masses. Fortunately, there appears to be a calculable correlation between mass and light-curve width, so they can still fill the 'standard candle' role, and research based on them is probably still valid. (I took data for the paper, but am not an author.)"

Comments

Re:Reassuring

[i kan reed]

If you don't end your paper with "Further research is needed" a secret cabal of other researchers also applying for grants in your field come into your bedroom at night, drag you off, and tie you to a post in the beam of the LHC.

so how far off is this?

[Charliemopps]

Our entire understanding of how the universe works is largly based on these measurements... so this is potentially a very big deal. Any Astrophysicists around that can give us an idea of how big a difference this will make?

Re:so how far off is this?

[Shag]

I don't think it's going to make a difference. In fact, I'm not quite sure whether the dark energy research that got the Nobel was strictly limited to type Ia supernovae - it was before my time, and since they were using high-Z (very distant) supernovae, they might have wanted more massive type II ones, or something.

For about a decade, people have accepted that some SNe Ia are "over the limit" (under arrest!) and have developed "double-degenerate" models of colliding white dwarf stars. As sky surveys discover more and more, it's started to become apparent that there are also some "under the limit." This project has studied hundreds of supernovae over the last decade, and looked pretty closely at how they evolve over time. The reassuring part of the paper is that even though these supernovae are nowhere near all the same mass blowing up every time, they're still within a reasonably sensible range (0.9 to 1.4 solar masses) and that by watching what brightness they reach at their peaks, and how quickly they decline in brightness, and looking at their spectral curves (all of which are among the things that this particular collaboration looks at), astrophysicists can calculate their masses, and thus make any necessary adjustments to compensate for that. And by the standards of astrophysicists and cosmologists, the math required to "standardize" progenitors of different masses is probably considered "easy." Of course, these are the same people who think "nearby" means 0.4-1.0 billion light years away...

Disclaimer: I am not an astrophysicist of any kind. I got involved in astronomy a decade ago, and took a few classes 5 years ago, but my roles are overwhelmingly technical or operations, and when it comes to science, I am always the "village idiot" surrounded by PhD's. I'm not the guy who'll give a lecture about what the telescope's pointing at - I'm the guy who'll fix the telescope so it points at it in the first place. I'll take data - in this particular case, over a 10-year project, I'll probably rank #1 or #2 in terms of amount of time spent taking data - but I don't do the analysis or write the papers. My background was in things like systems administration, spamfighting, web development, etc., as one would expect of someone with my user number here.

Re:so how far off is this?

[boristhespider]

" In fact, I'm not quite sure whether the dark energy research that got the Nobel was strictly limited to type Ia supernovae..."

No, they were definitely intended to be SN1a.

Riess et al.: http://arxiv.org/abs/astro-ph/...
"We present observations of 10 type Ia supernovae (SNe Ia)..."

Perlmutter et al.: http://uk.arxiv.org/abs/astro-...
"...All SN peak magnitudes are standardized using a SN Ia lightcurve width-luminosity relation..."

The reason is that SN1a can be standardised -- although that's an empirical (i.e. phenomenological) relationship rather than a theoretical one, it seems to be basically robust, as this paper has demonstrated -- and therefore used as standard candles. Other types of supernovae can not be used in the same way; one cannot necessarily correlate a (corrected) brightness against a (corrected) redshift.

This doesn't say that samples aren't contaminated by supernovae that aren't actually Type 1a (and a few years back an explanation for tension between the so-called "Gold sample" and other datasets was that it may have been more contaminated), but the intention is to only look at Type 1as.

I'd also argue that they weren't particularly high redshift, but then for me a redshift of 3 or 4 is very much low redshift. Come to that, redshifts of 300 are low redshift.

This is the Phillips relation - known for 20 years

[Anonymous Coward]

Mmm... press release spin.

There is good science here, but it is being heavily spun. The relation between light curve width and how bright SN are has been known since at least 1993 (Phillips, M., 1993, ApJ 413, L105). This was corrected for even in the original work that won the Nobel prize. So, the 'they aren't quite so standard candles' has been known for 20 years - what they are is 'standardizable' candles.

What is interesting about this work is that the SNFactory has tried to find a link between the Phillips relation and physical properties of the explosion - in this case the ejected Nickel 56 mass. And that is very interesting from a SN physics perspective, if not really so much from the 'doing cosmology with Type Ia Supernovae' one.

Re:Reassuring

[Ol Olsoc]

I thought "Further research is needed" was code for "we have seventeen more papers in the works, all of which will cite this one and each of the others, thus inflating our g-index and h-index numbers."

No, tha's the difference between researchers and teh political types. Research only raises more questions, and politicians know all the answers before any research is done.

Re:This is the Phillips relation - known for 20 ye

[Anonymous Coward]

Hello! I read the same press release, but I don't see the SuperNova Factory taking credit for discovering the relationship between light curve width and luminosity (the Phillips relation, which is indeed well-known, and made the discovery of Dark Energy with Type Ia supernovae possible). So if there is any spin here, it's not on that axis. The press release even acknowledges standardization to about 10% in distance.

Rather, the paper is about explaining this relationship in terms of the underlying physics as you point out. Interestingly, instead of the driver being the amount of radioactive nickel synthesized during the explosion (which is a good candidate for your main lever arm if you believe that total mass of Type Ia progenitors was fixed at/near the Chandrasekhar mass), it is the total mass itself that drives the relationship. It means that the conventional wisdom that the Chandrasekhar mass is somehow important needs another look.

And indeed, this paper *is* interesting also from the "doing cosmology with Type Ia supernovae" perspective. We don't know very well what kinds of white dwarfs and star systems can give us Type Ia supernovae. If there are many different channels, or a continuum of channels, or the distribution of masses of white dwarfs that make Type Ia supernovae somehow evolves with redshift, we need to understand that.

Re:This is the Phillips relation - known for 20 ye

[Shag]

AC1:

The relation between light curve width and how bright SN are has been known since at least 1993 (Phillips, M., 1993, ApJ 413, L105). This was corrected for even in the original work that won the Nobel prize. So, the 'they aren't quite so standard candles' has been known for 20 years - what they are is 'standardizable' candles.

AC2:

I don't see the SuperNova Factory taking credit for discovering the relationship between light curve width and luminosity (the Phillips relation, which is indeed well-known, and made the discovery of Dark Energy with Type Ia supernovae possible).

Well... the Phillips Relationship is "well-known" in much the same way that these supernovae are "nearby" - to the people in that specific very narrow field of expertise. Yes, Wikipedia has an article on it, but I'd expect it to be unknown to the average adult walking down the street, the average amateur astronomer, the average Jeopardy contestant, the average undergraduate or first-year graduate astronomy student, or even the average science popularizer who isn't specifically dealing with supernovae. Just last month, I overheard one long-time amateur astronomer still telling tourists at the Mauna Kea Visitor Information Station that all SNe Ia are the same mass and brightness!

But anyway, as the 2nd AC said, the newer/more interesting bit is the relationship to progenitor mass, and the continued trend toward SNe Ia coming from diverse progenitors - i.e. the more we look, the more "exceptions to the rule" we find. We're already to the point where it looks like most SNe Ia aren't from single, Chandrasekhar-mass progenitors as was long thought to be the "norm," and the paper discusses some models for progenitors of varying masses that meet with varying degrees of success in attempting to match the observational results. I suspect the computational / theoretical / modeling folks will also have fun with it all.

Re:This is the Phillips relation - known for 20 ye

[Anonymous Coward]

Hi all, I'm the lead author on the paper who got linked here by a colleague. I agree the Phillips relation is very well-known and well-established *empirically*, and that it is not endangered, nor is my boss going to have to give back his Nobel Prize because we found this out. But understanding this well-known empirical relation theoretically, from first principles, is a real challenge, and it's part of what's preventing us from further improving their performance as "standard candles". We need to continue to refine our understanding of SNe Ia to go to the *next* level in SN Ia cosmology, where we try to figure out what particle or field is responsible for causing the acceleration in the Universe's expansion. So understanding the physics of the supernova explosions is absolutely on the critical path to doing cosmology. The alternative approach is to just accumulate ten million SNe and self-calibrate, which is what LSST is plannign to do. But that will certainly take a lot of time on a half-billion-dollar telescope, to the extent that it's possible without taking spectra of each SN.

There are two main results here. One is the discovery that the ejected mass may be what's driving the Phillips relation, which was suggested before by theorists but never conclusively shown (what about asymmetries? 56Ni mass? weird ejecta density profiles?). The other is simply the fact that a large fraction of normal, non-peculiar SNe Ia may well not be Chandrasekhar-mass, thereby giving us a *lot* of information about how they blew up, and possibly how they evolved to the point of blowing up as well. So it touches on cosmology, but also on stellar evolution, chemical evolution of galaxies, basically every area of astrophysics.

Part of why the Chandrasekhar-mass picture lasted this long is because people didn't think you could standardize non-Chandrasekhar-mass explosions. But you can, and the differences in ejected mass may in fact be what allow us to do that.

(Ah, I see another anon has posted the same thing below. +2!)

Best,
RS